Automation can be attended where the technology is integrated with and dependent on human interaction, or unattended, where the technology can perform complete tasks without any human intervention. Automation also can be rule-based, which is particularly effective when deployed to take over repetitive, manual, large volume processes; or cognitive, where it is deployed to take on non-routine tasks that require some judgment.
Automation use cases
Process automation is a journey with standard, intermediate and advanced stages. Duke provided examples of how this journey might look in a healthcare provider organization.
One example related to a standard use case focused on rules-based, routine business processes. Providers have been using forms of standard automation for years to review claims. The technology and software, like RPA, gets more sophisticated every year, so now an automated program can gather real time claims data from payer portals and determine if a claim will be paid or denied, and then route that information to a human knowledge worker for appropriate next steps. Providers “get significantly more detailed information, and by knowing more, they can make their FTEs more effective,” Duke said. RPA utilization in some instances can replace up to 70 percent of a manual process.
The intermediate stage incorporates technology to automate tasks that are more complex. A practical application for healthcare providers is obtaining authorization for treatment. What previously required back-and-forth phone conversations between provider and health plan might be streamlined in part by an automated program, operating 24/7. Duke said this automation “helps with bottom line revenue, because if a provider has a tool in place that can obtain those authorizations, we should see a reduction in authorization denials.” The tool can also provide valuable information for the provider about why authorizations are denied in real time, which the provider can incorporate into its training and processes.
The advanced stage of intelligent automation incorporates machine learning and artificial intelligence to initiate actions that are more sophisticated, such as building a case to appeal a claim denial. If a claim is denied, an automated program could grab the electronic medical record (EMR) related to the patient and automatically submit it to the insurance plan. It also could deliver the EMR to a provider employee to approve before submission to a health plan. Either way, the provider is saving time and money by not needing an FTE to do the manual work of finding and forwarding an EMR.
Duke said there are hundreds of processes a provider can automate in some way so it is important for the provider to prioritize the most valuable tasks to consider for automation. He emphasized that activities concerning the revenue cycle and accounts payable are two areas where a provider should receive a high ROI using automation.
Duke walked through the process of how a provider would investigate and introduce intelligent automation to support revenue cycle activities. The first step is to make sure that executives, management, and staff – “from scheduling to claims adjudication” – are interviewed to identify frustrations, repetitive tasks and suggestions for automating tasks.
Next, the provider needs to look at the list of tasks that might be automated and start process mapping – watch how people currently do this work, determine if the task can be automated and put a value on that automation. One example, according to Duke, would be looking at denial outcomes or accounts receivable performance – tasks that have process-specific outcomes where an ROI can be determined if the provider automates the process.
Providers should also be prepared to engage subject matter expertise outside of the organization to see what has worked for other providers. Finally, when the provider starts to implement any automated solution, it should be prepared to balance industry and local best practices. Duke noted, “There are some things where the optimal result might not match up with the leading practice in the industry, but it might be the best you can do given some of the constraints that you have.” A key example of this is whether the provider decides to manage a new type of automation in-house or provide it as a service that a vendor builds and manages.
Garcia noted that as providers look to implement automation, they face common challenges:
- Common goals. While different departments at a healthcare provider may have different automation needs, the provider should strive to make sure the overall automation strategy supports the organization’s goals.
- Use case identification. A provider needs to filter, prioritize and score use cases to focus on the ones that provide the most value if automated.
- Rule-based or cognitive. The provider needs to be aware if it can support more sophisticated, cognitive automation or if it should focus on simpler, though still valuable, rules-based automation solutions.
- Business process mapping. If the provider has already used business process mapping, deploying that skill set to support intelligent automation will be easier.
- Automation anxiety. Even if the provider successfully does the work to deploy automation, it has to communicate effectively to all staff what is being done and how letting automation take over more mundane tasks will free up staff to do higher value activity.
- ROI and measurement. A provider needs to anticipate how ROI will be measured as the automation is put in place.
Duke emphasized, “A strong ROI gives the organization a lot of confidence that this process automation journey is worth it.”
A recording of the webinar is available here.